Acoustic Modelling

A simple multiband approach for solving frequency dependent problems in numerical time domain methods

by Jonathan Sheaffer

Sheaffer, J, Fazenda, BM, Murphy, DT and Angus, JAS 2011 , in: Forum Acusticum, 26 June - 1 July 2011, Aalborg, Denmark.

With the rapid growth of computational power and recent advances in GP-GPU technology, numerical time domain methods... more With the rapid growth of computational power and recent advances in GP-GPU technology, numerical time domain methods are becoming increasingly popular for room acoustics applications due to their accuracy, simplicity and ease of implementation. However, in order to model realistic spaces one should consider boundary conditions and source directivity functions as empirically measured frequency dependent quantities. Previously suggested methods rely on performing time domain convolution or employing recursive filters at the boundaries of the domain. Although shown to be highly accurate, these formulations normally involve complex implementations which not only reduce the attractiveness of using such methods, but may also result in computationally expensive algorithms. In this paper we examine a straightforward approach for solving such frequency dependent problems at the expense of being able to run a single broadband simulation. Using the finite difference time domain (FDTD) method, a number of band-limited frequency independent simulations are initiated in intervals depending on the availability of empirical impedance and directivity data and the desired spectral resolution. Generated impulse responses are filtered according to their respective frequency bandwidths and summed to produce a single frequency dependent impulse response. Intermediate values are automatically interpolated based on the characteristics of the chosen post processing filters. Results are analysed and validated and agreement with theoretical models is shown.

On the Simulation of Porous Media by Means of Finite Difference Time Domain Algorithms

by Rubén Picó

Time-domain formulations for sound propagation in rigid-frame porous media. The equations proposed by Zwicker and... more Time-domain formulations for sound propagation in rigid-frame porous media. The equations proposed by Zwicker and Kosten (ZK) in their model [Zwicker] are appropriated to perform numerical simulations in the time domain with the FDTD method. The stability conditions of FDTD schemes based on the ZK equations are investigated in this paper. The Von Neumann method is used to assure the stability response of each Fourier component of sound fields of the discrete scheme. The insertion loss of a sandwich-type configuration with porous media is analyzed as a numerical application.

Time Domain Simulation of Sound Diffusers Using Finite-Difference Schemes

by Rubén Picó

Since the invention of sound diffusers three decades ago a substantial effort has been made to predict the acousticmore Since the invention of sound diffusers three decades ago a substantial effort has been made to predict the acoustic
behaviour of these structures, for auralisation and prediction purposes as well as in response to the large
costs inherent in anechoic measurements. Volumetric methods such as Finite Element Methods (FEM) or the
Finite Difference Time Domain method (FDTD) are not often used, due to their large computational cost. However
Near Field to Far Field Transformations (NFFFT) can overcome that problem. The main advantages of the
FDTD method are that a single calculation is sufficient to study a wide frequency band, and that the time domain
behaviour of the reflected sound can be directly inspected. In this paper we present a comparison between the
prediction techniques commented above in the context of sound diffusers, paying special attention to the FDTD
method. Having demonstrated that the FDTD method can generate results comparable to more established techniques,
early results concerning the modelled performance of diffusers in the time domain (‘time spreading’) are
reported, opening a new field of research.

Prediction of the Random-Incidence Scattering Coefficient Using a FDTD Scheme

by Rubén Picó

This paper is focused on the evaluation of the scattering coefficient of sound diffusers which are based on themore This paper is focused on the evaluation of the scattering coefficient of sound diffusers which are based on the
incoherency of diffusely reflected sound. A new approach for predicting the scattering coefficient is proposed; the
method is based on a finite difference time domain (FDTD) scheme. Two established scattering coefficient measurement
methods, proposed by Mommertz and Vorländer [1], are simulated; these correspond to measurements
in a reverberant chamber, and free field. The results are also compared to those obtained in a previous paper [2],
wherein it was demonstrated that FDTD schemes can be used to predict polar responses. These free field polar
responses are used to find the correlation scattering coefficient, which in turn is used to validate the free field
case. In modelling the reverberant chamber method, 2D simulations have been used to reduce computation time;
hence it is necessary to derive a diffuse field formulation for a 2D reverberation chamber, which is presented. In
this case, 1:5 scale experimental data is used for validation.

"FDTD/K-DWM Simulation of 3D Room Acoustics on General Purpose Graphics Hardware using Compute Unified Device Architecture (CUDA)

by Jonathan Sheaffer

in Proc. Institute of Acoustics Vol. 32 Pt. 5 2010.

The growing demand for reliable prediction of sound fields in rooms have resulted in adaptation of various approaches... more The growing demand for reliable prediction of sound fields in rooms have resulted in adaptation of various approaches for physical modeling, including the Finite Difference Time Domain (FDTD) and the Digital Waveguide Mesh (DWM). Whilst considered versatile and attractive methods, they suffer from dispersion errors that increase with frequency and vary with direction of propagation, thus imposing a high frequency calculation limit. Attempts have been made to reduce such errors by considering different mesh topologies, by spatial interpolation, or by simply oversampling the grid. As the latter approach is computationally expensive, its application to three-dimensional problems has often been avoided. In this paper, we propose an implementation of the FDTD on general purpose graphics hardware, allowing for high sampling rates whilst maintaining reasonable calculation times. Dispersion errors are consequently reduced and the high frequency limit is increased. A range of graphics processors are evaluated and compared with traditional CPUs in terms of accuracy, calculation time and memory requirements.

Bifurcation Analysis of a Vocal Fold Model Coupled to Resonators

by Haralambos Hatzikirou

Voice Instabilities Due to Source-Tract Interactions

by Haralambos Hatzikirou

Vocal production mechanisms in a non-human primate: Morphological data and a model

by Haralambos Hatzikirou

Multiple discontinuities in nonhuman vocal tracts - A response to Lieberman

by Haralambos Hatzikirou